For AI builders
Your agent built it. Now find out if it works
Paste the URL of the prototype your agent just built, or ask the agent itself. Fiuto writes the study, runs it with AI respondents and real people, and reads back what to change before the next build.Free to start, in minutes, from Claude Code, Cursor, Lovable, v0, Replit, Bolt, or the web app.
How it works
Fiuto runs your full research loop
Paste a URL and Fiuto takes it from there. No research experience needed.
A working prototype is not the answer yet
Getting from idea to a working web prototype is suddenly the easy part. The risky bit is sending the agent back for another build before anyone outside the project has tried the thing. A polished demo can still leave a first-time user looking for the wrong button, missing the point of the screen, or abandoning the task you thought was obvious. You are close to the build. They are not.
Fiuto gives someone outside the project a real task on your prototype and room to explain what happened. You get the behaviour and the reason together: where they went, where they stopped, and what made sense to them. Then you hand the findings back to the agent and build the change their experience earned, instead of another confident guess.
What to test before the next build
Start from the question in front of you. Each one is a study Fiuto can draft from your prototype in minutes.
Can someone finish the first task?
Give a new person the job the prototype is meant to make easy, screen recorded, and see whether they get through it and where they lose the thread. Explore prototype testing.
Do they know where to click first?
When the first move matters, show the screen and ask where they would go. A first-click test catches the wrong turn before you build the rest of the path. Explore first-click testing.
Which direction should survive?
Show the alternatives you are debating, ask people to choose, then ask why. A preference test turns taste into a decision you can inspect. Explore preference testing.
Does the landing page say what you think it says?
People see the page for five seconds and tell you what it does and who it is for. If they get it wrong, the copy goes back to the agent, not the product. Test your landing page.
What you get as a builder
The whole research loop, run for you, from the place you already build.
A study drafted from your prototype, not a blank builder
Paste the URL and say what you need to learn. Fiuto reads the prototype and drafts the whole study, task, screening, and follow-up questions, for you to review. You never have to know which method to pick.
The full loop from the agent you build in
Fiuto's MCP server creates, launches, and reads back studies from Claude Code, Cursor, Codex, Lovable, v0, Replit, and Bolt, on every plan including Free. Ask the agent to test what it just built, in the same session.
AI respondents go first
A cast of AI respondents works through the prototype before anyone else, so you get a first read in minutes and catch a broken task before real people see it. Up to two of the cast can be AI experts in conversion, usability, copy or SEO, on every plan including Free.
Real people, recruited or your own
Describe the audience you need and Fiuto recruits it from Prolific's global panel, with demographic filters. Or share the link with your waitlist, community, or early customers. Runs are labelled by who answered.
Findings the agent can act on
Results come back as findings, a readout, and a deck, in the app or over MCP. Ask the agent to read them and propose the next change, so the evidence goes straight into the next build.
It remembers your product
Fiuto keeps a profile of your product, your audience, and what you are trying to learn, and carries it into every study it drafts. The second test is faster than the first, and the tenth is faster still.
Unlimited collaborators, no seats
Fiuto's price is per account. Share results, findings and decks with as many collaborators as you like on every plan, including Free. Nobody needs a licence to open them.
MCP for user research
Your data is yours. Stay in Claude Code
Ask the agent that built the prototype to test it. Create, launch, and read back studies over MCP from Claude Code, Cursor, Lovable, v0, Replit, Bolt, or Codex. Every result is yours, out as CSV or JSON through a per-study token, the API or MCP, on every plan including Free.
Create studies from
Claude Code
Cursor
Codex
Lovable
v0
Replit
BoltFree to start, upgrade when you need more
The free plan needs no card: every test method, 1 live study a month, 10 responses per study, and 600 AI credits a month, with the AI drafting, the AI respondents, and the findings all included. Core is $29 a month, flat per account, with 3 live studies a month, 100 responses per study, and 3,000 AI credits. Pro is $69 a month with 10 live studies, unlimited responses per study, and 12,000 AI credits.
Share a study, its results or its deck with as many collaborators as you like, on every plan including Free, and export the raw data as CSV or JSON on any of them. Upgrade, downgrade or cancel any time from the billing page, with no sales call at any step. Recruited participants are paid for on top of the plan, and the total is shown before you order.
Frequently asked questions
01Can I test a prototype built with Claude Code, Codex, Cursor, Lovable, v0, Replit, or Bolt?
Yes. If your agent made a web prototype, paste its URL into Fiuto or ask the agent to create the study over MCP. Fiuto drafts the study, runs it with AI respondents and real people, and reads back what they did and said before the next build.
02Can I run the whole thing from my coding agent?
Yes. Connect an MCP-ready coding agent to Fiuto, then ask it to create and launch a study from the same session where you are building, and to read the results back when they are in. You can also start in the web app and use the agent later. Both routes meet in the same study.
03Does Fiuto recruit test participants?
Yes. Describe the audience you need and Fiuto recruits it from Prolific's global panel, with demographic filters. You can also share the study link with your own waitlist, community, or customers, and the two kinds of run stay labelled separately.
04What are AI respondents?
A cast of AI respondents that works through your study before any real person does. They give you a first read in minutes, and catch a task that is broken or unclear before you spend real responses on it. They are a rehearsal, not a substitute: the real answers come from real people.
05What does prototype tracking add?
For an agent-built browser prototype whose code your agent can edit, tracking adds what respondents did inside the prototype alongside their study answers. It is optional. For a static screen, a first-click or preference test often answers the question more directly.
06Do I need research experience?
No. Paste the URL and say what you need to learn. Fiuto writes the study, runs it, and reads the results back into findings and a deck. You review each step before it launches, and you can change anything.
07Is it free to start?
Yes. The free plan includes every method, 1 live study a month, 10 responses per study, and 600 AI credits a month. No card, no time limit.
08Does Fiuto run AI interviews?
AI interviews are an upcoming feature, in development now, and this page will list them the day they are out of alpha. Until then Fiuto runs unmoderated studies across 18 methods, with AI respondents first and real people after.
09Can I change or cancel my plan without contacting anyone?
Yes. You can upgrade, downgrade, or cancel your subscription at any time from the billing page in the app. There is no sales conversation at any step, from the free plan onwards.